High speed water monitoring systems based on Digital Holographic Microscopy

Using Digital Holographic Microscope (DHM) technology - as it alleviates the small depth of focus constraint of the conventional microscopes - the fast screening of large fluid sample volumes becomes achievable. We have developed and built automatic DHM technology based fluid monitoring devices which are able to determine the biological content of water samples. These devices are capable for the detection, segmentation and imaging of sparse, freely floating objects in hundred times larger volumes than counterparts applying conventional microscopy approches. The aimed automatic pattern recognition and object classification requires high quality image reconstructions from the acquired hologram of the objects. To fulfill this task, the precise 3D position determination of the objects is required. This means, proper auto-focusing and object segmentation algorithms has to be applied. We outline the introduced algorithms, which are based on the special, coherent imaging properties of the applied in-line holographic systems. Thanks to the rapid development of the digital computer technology and the application of parallel computing implementations the numercal evaluation of the recorded holograms can be fulfilled close to real time. Using stream processors (graphics processing unit, GPU) it is possible to increase the algorithm speed considerably, without any perceptible reconstruction accuracy loss.